Show HN: Corgi, a helpful {agent, terminal} multiplexer

Who Should Read

Developers running several coding agents like Claude Code and Codex in the terminal who struggle to track their status

Core Mechanics

• Agent status shown by color
  • It distinguishes working, needs you, finished and idle.
  • You can see at once which agent needs you.
• Supports Claude Code, Codex, Pi and OpenCode
  • Ctrl-T opens a new agent tab or jumps into existing work.
  • It uses your existing accounts.
• Keeps running in the background like tmux
  • Agents and processes like a dev server survive closing the terminal.
  • Run corgi to get back in.
• Tracks tokens, changed files and resources
  • top for your tokens shows where tokens went.
  • You can scroll through every file an agent touched and see per-workspace resource usage.
• Worktree management and SSH image paste
  • It organizes workspaces for you.
  • ⌘V image paste works over SSH but requires Ghostty Nightly.
• Agents can drive Corgi
  • You can control Corgi from any terminal.
  • corgi --skill teaches your agents how to use it.

Community Reactions

• herdr users called it a v2
  • Several commenters said they would switch from herdr.
  • One user who used it for a month said it is a good cmux replacement.
• Multi-machine connection is a weak spot, some said
  • One commenter said it doesn't connect seamlessly across machines like herdr, so they can't switch.
  • Another said multiple local connections are core for them.
• Users reported bugs after trying it
  • One reported repaint and flicker bugs.
  • Others noted clunky theming and worried about huge TS files in the source.
• Pushback on install method and name
  • One asked for brew instead of curl | bash.
  • Another pointed out a YC company named Corgi already exists.

How to Apply

• Install and run two agents side by side
  • Run the install command on macOS, Linux or WSL.
  • Open Claude Code and Codex in separate tabs and check the status indicators.
• Compare with your tmux agent setup
  • Run the same task in tmux and Corgi and compare notifications and status visibility.
  • Close the terminal and reattach to confirm background persistence.
• Find costly agents via token tracking
  • Open the top for your tokens view while running several agents.
  • Pick the token-heavy tasks and shrink their prompts or scope.

Try It

# Paste into your terminal (macOS, Linux, WSL)

curl -fsSL https://corgi.computer/install.sh | bash

# Run after install
corgi

# Next: open an agent tab with Ctrl-T and teach your agent to use Corgi with corgi --skill

Terminology

Related Papers

• Ask HN: If the models are so good, why doesn't OpenAI just start 100 companies?

  HN answers: code is easy, knowing what to build isn't, so they sell tokens instead

• OpenShell – A safe, private runtime for fleets of autonomous AI agents

  OpenShell uses policy files to control agent file, network, and credential access at kernel level.

• Show HN: Omnara – Open-source alternative to Claude Managed Agents

  Open-source managed agent platform runs agents on any model and machine, self-hostable

• Rust off Rails: The opinionated web stack for the AI era

  Sixty Markdown files tell coding agents how to build a Rust web app, with no framework code.

• LayerFS: The filesystem for AI coding agents

  Mount an ephemeral workspace per agent, commit only the changes

• Launch HN: Alter (YC S25) – Access control for multi-user agents

  Alter gives each agent its own identity, policy, and audit trail instead of shared API keys.

• Using Claude to reduce the attack surface of code

  Claude scanned 25 open-source projects and found 493 needlessly reachable code issues.

• WhatsApp's help pages hid LLM instructions: dont recommend a regulator

  WhatsApp help pages hid 102 AI-only instructions, including not recommending a data regulator

Related Resources

• https://corgi.computer/
• https://exawatt.ai
• https://nocurlbash.com/#en
• https://www.ycombinator.com/companies/corgi-insurance